Search

We introduce Single-cell Nanopore Spatial Transcriptomics (scNaST), a software suite to facilitate the analysis of spatial gene expression from second- and third-generation sequencing, allowing to generate a full-length near-single-cell transcriptional landscape of the tissue microenvironment. Taking advantage of the Visium Spatial platform, we adapted a strategy recently developed to assign barcodes to long-read single-cell sequencing data for spatial capture technology. Here, we demonstrate our workflow using four short axis sections of the mouse heart following myocardial infarction. We constructed a de novo transcriptome using long-read data, and successfully assigned 19,794 transcript isoforms in total, including clinically-relevant, but yet uncharacterized modes of transcription, such as intron retention or antisense overlapping transcription. We showed a higher transcriptome complexity in the healthy regions, and identified intron retention as a mode of transcription associated with the infarct area. Our data revealed a clear regional isoform switching among differentially used transcripts for genes involved in cardiac muscle contraction and tissue morphogenesis. Molecular signatures involved in cardiac remodeling integrated with morphological context may support the development of new therapeutics towards the treatment of heart failure and the reduction of cardiac complications.

Our understanding of the transition from physiological to pathological cardiac hypertrophy remains elusive and largely based on reductionist hypotheses. Here, we profiled the translatomes of 15 mouse hearts to provide a molecular blueprint of altered gene networks in early cardiac remodeling. Using co-expression analysis, we showed how sub-networks are orchestrated into functional modules associated with pathological phenotypes. We discovered unappreciated hub genes, many undocumented for their role in cardiac hypertrophy, and genes in the transcriptional network that were rewired in the translational network, and associated with semantically different subsets of enriched functional terms, such as Fam210a, a novel musculoskeletal modulator, or Psmd12, implicated in protein quality control. Using their correlation structure, we found that transcriptome networks are only partially reproducible at the translatome level, providing further evidence of post-transcriptional control at the level of translation. Our results provide novel insights into the complexity of the organization of in vivo cardiac regulatory networks.